DYNAMIC FLEET MANAGEMENT Concepts, Systems, Algorithms & Case Studies

OPERATIONS RESEARCH/COMPUTER SCIENCE INTERFACES SERIES

Professor Ramesh Sharda Prof. Dr. Stefan Voß Oklahoma State University Universität Hamburg

Greenberg /A Computer-Assisted Analysis System for Mathematical Programming Models and Solutions:

A User’s Guide for ANALYZE Greenberg / Modeling by Object-Driven Linear Elemental Relations: A Users Guide for MODLER Brown & Scherer / Intelligent Scheduling Systems Nash & Sofer / The Impact of Emerging Technologies on Computer Science & Operations Research Barth / Logic-Based 0-1 Constraint Programming Jones / Visualization and Optimization Barr, Helgason & Kennington / Interfaces in Computer Science & Operations Research: Advances in

Metaheuristics, Optimization, & Stochastic Modeling Technologies Ellacott, Mason & Anderson / Mathematics of Neural Networks: Models, Algorithms & Applications Woodruff / Advances in Computational & Stochastic Optimization, Logic Programming, and Heuristic

Search Klein / Scheduling of Resource-Constrained Projects Bierwirth / Adaptive Search and the Management of Logistics Systems Laguna & González-Velarde / Computing Tools for Modeling, Optimization and Simulation Stilman / Linguistic Geometry: From Search to Construction Sakawa / Genetic Algorithms and Fuzzy Multiobjective Optimization Ribeiro & Hansen / Essays and Surveys in Metaheuristics Holsapple, Jacob & Rao / Business Modelling: Multidisciplinary Approaches — Economics, Operational

and Information Systems Perspectives Sleezer, Wentling & Cude/Human Resource Development And Information Technology: Making Global

Connections Voß & Woodruff / Optimization Software Class Libraries Upadhyaya et al / Mobile Computing: Implementing Pervasive Information and Communications

Technologies Reeves & Rowe / Genetic Algorithms—Principles and Perspectives: A Guide to GA Theory Bhargava & Ye / Computational Modeling And Problem Solving In The Networked World: Interfaces in

Computer Science & Operations Research Woodruff / Network Interdiction And Stochastic Integer Programming Anandalingam & Raghavan / Telecommunications Network Design And Management Laguna & Martí / Scatter Search: Methodology And Implementations In C Gosavi/ Simulation-Based Optimization: Parametric Optimization Techniques and Reinforcement

Learning Koutsoukis & Mitra / Decision Modelling And Information Systems: The Information Value Chain Milano / Constraint And Integer Programming: Toward a Unified Methodology Wilson & Nuzzolo / Schedule-Based Dynamic Transit Modeling: Theory and Applications Golden, Raghavan & Wasil / The Next Wave in Computing, Optimization, And Decision Technologies Rego & Alidaee/ Metaheuristics Optimization via Memory and Evolution: Tabu Search and Scatter

Search Kitamura & Kuwahara / Simulation Approaches in Transportation Analysis: Recent Advances and

Challenges Ibaraki, Nonobe & Yagiura / Metaheuristics: Progress as Real Problem Solvers Golumbic & Hartman / Graph Theory, Combinatorics, and Algorithms: Interdisciplinary Applications Raghavan & Anandalingam / Telecommunications Planning: Innovations in Pricing, Network Design and

Management Mattfeld / The Management of Transshipment Terminals: Decision Support for Terminal Operations in

Finished Vehicle Supply Chains Alba & Martí/ Metaheuristic Procedures for Training Neural Networks Alt, Fu & Golden/ Perspectives in Operations Research: Papers in honor of Saul Gass’ 80th Birthday Baker et al/ Extending the Horizons: Adv. In Computing, Optimization, and Dec. Technologies

DYNAMIC FLEET MANAGEMENT Concepts, Systems, Algorithms & Case Studies

Edited by Vasileios Zeimpekis Christos D. Tarantilis George M. Giaglis Ioannis Minis

Vasileios Zeimpekis Christos Tarantilis Athens University of Economics & Business Athens University of Economics & Business Athens, Greece Athens, Greece George M. Giaglis Ioannis Minis Athens University of Economics & Business University of Aegean Athens, Greece Chios, Greece Series Editors: Ramesh Sharda Stefan Voß Oklahoma State University Universität Hamburg Stillwater, OK, USA Hamburg, Germany

Library of Congress Control Number: 2007924349

ISBN-13: 978-0-387-71721-0 e -ISBN-13: 978-0-387-71722-7 Printed on acid-free paper. © 2007 by Springer Science+Business Media, LLC All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now know or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks and similar terms, even if the are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. 9 8 7 6 5 4 3 2 1 springer.com

TABLE OF CONTENTS

Acknowledgments

1. Planned route optimization for real-time vehicle routing

Soumia Ichoua, Michel Gendreau and Jean-Yves Potvin 1

2. Classification of dynamic vehicle routing systems

Allan Larsen, Oli B.G. Madsen and Marius M. Solomon 19

3. Dynamic and stochastic vehicle routing in practice

Truls Flatberg, Geir Hasle, Oddvar Kloster, Eivind J. Nilssen and Atle Riise

41

4. A parallelizable and approximate dynamic programming-based

dynamic fleet management model with random travel times and multiple vehicle types Huseyin Topaloglu

65

5. Integrated model for the dynamic on-demand air

transportation operations Yufeng Yao, Özlem Ergun and Ellis Johnson

95

6. An intermodal time-dependent minimum cost path algorithm

Elaine Chang, Evangelos Floros and Athanasios Ziliaskopoulos 113

7. Real-time emergency response fleet deployment: concepts,

systems, simulation & case studies Ali Haghani and Saini Yang

133

8. Vehicle routing and scheduling models, simulation

and city logistics Jaime Barceló, Hanna Grzybowska and Sara Pardo

163

vii

xiii

Preface

vi Table of Contents

9. Dynamic management of a delayed delivery vehicle in a city logistics environment Vasileios Zeimpekis, Ioannis Minis, Kostas Mamassis

197

10. Real-time fleet management at eCourier Ltd

Andrea Attanasio, Jay Bregman, Gianpaolo Ghiani and Emanuele Manni

219

Index 239

and George M. Giaglis

PREFACE The challenges of contemporary fleet management are moving beyond cost efficiency towards superior customer service, agility, and responsiveness to requirements that vary at a time scale unthinkable even a decade ago. Over the last forty years classical methods of fleet management have addressed extensively the issue of cost efficiency by developing a priori routing plans in a wide spectrum of practical problems. However, the use of an initial plan, although necessary, is by no means sufficient to address events that are likely to occur during plan execution and significantly affect system per-formance. Typical examples of such evens are customer orders that arrive in real time and should be served by vehicles already on route, as well as disturbances intrinsic to urban environments, such as traffic delays, parking unavailability, and breakdowns. The ability to deal with such cases in a satis-factory manner is increasingly important to the competitiveness of logistics and transport related operations.

Dynamic fleet management refers to environments in which information is dynamically revealed to the decision maker. This information may not be known at the initial planning stage, and/or may change after the construction of the initial fleet routes during plan execution. In addition, there are significant cases in which no routing exists and the system responds to requests that arrive dynamically.

Methods that address the critical issues of dynamic fleet management may be implemented in practical systems by taking advantage of recent advances in satellite and mobile communication technologies. Specifically, satellite location identification systems that use the Global Positioning System (GPS) and terrestrial mobile communication systems, such as the General Packet Radio Service (GPRS) or Terrestrial Trunked Radio (TETRA), enable fleet operators to monitor the execution of a plan and to manage operations in real time, thus improving fleet performance.

This edited volume aims to highlight important advances in the emerging field of Dynamic Fleet Management. The fundamental problem of real time vehicle routing is defined and solution methods are presented and classified. Emphasis is also given to algorithmic approaches that are able to process dynamic information and produce solutions of acceptable quality for significant dynamic fleet management problems in almost real time. Finally, the volume includes case studies that address actual dynamic problems by combining systemic and algorithmic approaches.

The first three chapters survey important aspects of the dynamic vehicle routing problem.

viii Preface

Chapter 1 reviews and classifies solution methodologies that address real-time vehicle routing problems where customer requests are dynamically revealed over time. Each service request either has a combined pick-up and delivery location or only a single pick-up (or delivery) location. Different algorithmic methodologies are presented that handle the occurrence of new requests through the construction of the part of the route that has not yet been executed by the vehicle (i.e. planned route). To construct the planned routes, adaptations of methods originally employed to solve the static problem are presented. Issues of diverting a vehicle away from its current planned destination to serve a new request that has just occurred are then discussed. Solution methodologies that anticipate future requests to efficiently satisfy future demands are also reviewed. Chapter 1 concludes by proposing future directions for research, such as the development of solutions approaches for handling the occurrence of vehicle breakdowns and unexpected congestion, formal modeling frameworks that integrate the uncertainty associated with future requests, as well as other theoretical and practical issues of research.

Chapter 2 discusses important characteristics and properties of the dynamic vehicle routing problem from a temporal point of view. Differences between the static and the dynamic vehicle routing problems are also demonstrated by analyzing critical issues from previous published papers. The importance of measuring the performance of a dynamic vehicle routing system is then highlighted, and measures for dynamism in systems with and without time windows are discussed. Methods for evaluating the performance of on-line routing algorithms are presented and important issues to include in the system objective are reported. A three-echelon classification of dynamic vehicle routing systems is also proposed based on a) their degree of dynamism and b) the objective of the system. Finally, Chapter 2 emphasizes the significance of considering the volume and the temporal composition of immediate requests along with the system objective when developing an algorithmic methodology for a dynamic vehicle routing system.

Chapter 3 discusses the experience gained in practical issues in stochastic and dynamic routing in the context of developing a VRP solver at a Norwegian research institute. First, a review of the literature on dynamic and stochastic vehicle routing problem (DSVRP) is presented. To illustrate the need for dynamic and stochastic models in real world applications, two examples involving transportation of goods and persons, respectively, are demons-trated. Modelling and formal description of DSVRPs is also proposed to create the platform upon which new computational methods will be tested and evaluated. Chapter 3 also considers the context in which a VRP solver operates and proposes solution approaches based on scenario generation. The way to exploit dynamic events to produce more robust plans to the VRP is discussed, as well as the role of generating statistical knowledge of events

Preface ix

automatically from past experience. Chapter 3 concludes with suggestions for further research in these areas.

Chapters 4 to 6 provide algorithmic approaches of significant appli-cability to practical dynamic fleet management problems.

Chapter 4 proposes a dynamic programming-based approach to address a general fleet dispatching problem. In this problem the vehicles are dispatched to serve load requests, which arise randomly during each time period of a finite time horizon at different locations in a transportation network. The fleet comprises vehicles of multiple types. An additional critical (and practical) complication is that the travel times between the network nodes are random. The approach presented in this chapter uses modelling and methodological concepts from the deterministic travel time case to present a novel approach for random travel times. Major contributions of this chapter include the formulation of the problem as a dynamic program; the way of approximating the value function of the resulting subproblem for each time period by separable piecewise linear concave functions; the proof that the approximate subproblem is a min-cost network flow problem; the further decomposition of the latter to multiple problem instances by location, which can be solved in parallel. Furthermore, an updating method is employed to improve the value function approximations, and a comprehensive algorithm is proposed to obtain solutions of superior quality, as evidenced by the experimental results of three classes of problems included in this work. Chapter 4 concludes by proposing challenging new opportunities for research in the dynamic fleet management area, such as the introduction of load pick up and delivery windows when the load requests arrive randomly, as well as other unresolved practical issues.

Chapter 5 proposes a column generation-based approach to plan the operations of an on-demand air transport system. The problem consists of determining the fleet assignment, aircraft routing and crew pairing in an integrated fashion for a system that provides point-to-point service at customer request. The dynamic elements of the problem are twofold: i) The demand for service is not known in advance, and is dynamically received. ii) There are unscheduled maintenance requirements that are also raised dynamically. The proposed model is based on a three-day planning period within a rolling horizon setting. It uses a crew duty network and a fleet-station time line in order to embed the crew and aircraft information in the fleet assignment problem, while keeping the crew and aircraft separate during planning. In addition to the model, the major contributions of this chapter include: The use of column generation and identification of good pairings by solving special shortest path problems; the dynamic adjustment of the plan when new requests for service or unscheduled maintenance are revealed without relying on demand forecasts; the comparison between cases with fixed (immovable) requests, and cases in which (limited) freedom is

x Preface

given in satisfying a request. Experimental results from practical cases indicate the ability of the model and the proposed approach to deal effectively and on-time with the dynamic nature of demand requests in a realistic setting.

Chapter 6 addresses a transportation problem with time varying parameters. Specifically, it models and solves the problem of determining optimal paths in a transportation network with time dependent link costs and travel times. In addition, multiple modes of transport are considered, and the related transfer delays and costs are also time dependent and fully accounted for. The problem is solved to optimality by a minimum cost path algorithm that computes optimum path trees from all network nodes and feasible discrete departure times. The algorithm has been applied to intermodal routing in the case of hazardous materials transport. In this case the risks associated with mode-link combinations and transhipments are also time dependent, and the problem has been formulated in a way amenable to the proposed algorithm. It should be noted that due to its computational efficiency, the algorithm could be applied to dynamic problems that account for real time system changes.

Finally, Chapters 7 to 10 discuss real-life applications and case studies of dynamic vehicle routing and fleet management, demonstrating the appli-cability and practical significance of research in the area.

Chapter 7 introduces the need for real-time fleet management in emergency response situations. The authors propose an integrated emergency response fleet deployment system that embeds an optimization approach to assist dispatchers in assigning emergency vehicles to emergency calls, while having the capability to look ahead for future demands. The proposed system is tested and validated by means of a simulation model and a case study application in the area of Washington, DC. Moreover, a mathematical model for real time vehicle dispatching is presented and it is shown that its exact solution, minimizing the expected total wait over a large network, can be obtained with a short computation time.

Chapter 8 addresses the important area of City Logistics and reports on a DSS-based modelling framework aiming at supporting the design and evaluation of city logistics applications prior to their implementation. The decision support system draws on an underlying dynamic traffic simulation model that feeds a dynamic router and scheduler, which can then determine which vehicle to assign to new services as well as the new route for the selected vehicle. Further to the presentation of the proposed DSS, the chapter also discusses two case studies performed in the Italian cities of Lucca and Piacenza to illustrate how the system works in practice.

Chapter 9 also focuses on city logistics and discusses the design and implementation of a real-time fleet management system capable of rerouting

Preface xi

vehicles in real time when unforseen events, such as breakdowns or delayed vehicles that cannot meet future customer time windows, occur during urban freight distribution. The vehicle-mounted wireless communication sub-system monitors each vehicle through GPS-based positioning that is reported to the dispatch centre via GPRS. The dispatch centre utilizes this information to monitor the fleet, detect deviations from the initial distribution plan, and adjust the schedule accordingly by suggesting effective rerouting interventions. The chapter discusses the application of the system in one case study of a Greek 3PL operator, demonstrating the degree of customer service improvement that can be achieved through real time vehicle monitoring and rerouting.

Chapter 10 describes a real-time fleet management system designed and implemented for eCourier Ltd at London, UK. The chapter reports the overall system architecture, the main algorithms, the travel time forecasting procedure, and the job allocation heuristic used. The system is capable of monitoring courier location information and vehicle type, among other variables, in real-time. This information is fed to a set of algorithms that allocate each job to the most appropriate courier on the basis of road congestion and current fleet status, as well as individual courier efficiency. Courier location information is provided by GPS devices embedded into palmtop computers which are also used to provide directions to couriers. Results of system operation in real-life demonstrate its ability to reduce the requirements for human fleet management supervisors, to improve service and to increase courier efficiency.

ACKNOWLEDGMENTS

Preparing an edited volume is an exciting task based on collaboration and support by esteemed colleagues and co-workers. We owe gratitude to all authors who contributed their work on state-of-the-art methods and results related to dynamic fleet management.We would also like to express our appreciation to the chapter referees for their invaluable help in ensuring the quality standards of this volume. Specifically, we would like to thank: N. Altay, N. Ampazis, J. Barceló, E. Benavent, I. Benyahia, E. Chang, H.K. Chen, A. Corberan, L. Coslovich, G. Dounias, T. Fahle, G. Ghiani, A. Haghani, G. Hasle, J. Herrmann, S. Ichoua, G. Ioannou, B. Kallehauge, J. Q. Li, E. Manni, M. Montemanni, E. Mota, R. Nagi, R. Pesenti, J.-Y. Potvin, D. Pisinger, H. Psaraftis, M. Reimann, E. Taniguchi, P. Tsilingiris, B.W. Thomas, H. Topaloglu, S. Yang, Y. Yao, A. Ziliaskopoulos, and P. Zito. This volume would not be possible without the input, guidance and support of Gary Folven, editor-in-chief of the Operations Research stream in Springer Verlag and Carolyn Ford, editor assistant. Finally, many thanks are due to L. Amygdalou, G. Ninikas and T. Athanasopoulos for their support in editing and preparing the overall manuscript. V. Zeimpekis C.D. Tarantilis G.M. Giaglis I. Minis Athens and Chios, March 2007